Hacker News Reader

Comment by dijit | Hacker News Reader

> Jolla / Sailfish is a 13 year old project

Realistically building a production quality database takes 10 years. Building a production quality game engine takes 10 years.

They're building a mobile operating system and the hardware it runs on; that's harder and a moving target.

How long do you think it takes to build a supply chain of hardware that doesn't suck (if it takes 2 years to get moving: you need to start with hardware specs for 2 years from now) and an operating system that doesn't suck when you're also trying to catch up to a major duopoly cranking out devices at an unfathomable volume, with more money than most nation states?

Your standard is "succeed against Google and Apple within 13 years on a shoestring budget with no volume discounts." How can any project clear that bar?

What would you do?

View on HN

> Your standard is "succeed against Google and Apple within 13 years..."

Absolutely not. My standard is the many other AOSP-based ROMs communities and companies that were founded around them, having success within a few years - yes, they could lean on the ecosystem compatibility and didn't produce their own hardware, but maybe that's a more viable way to start?

"shoestring budget with no volume discounts" does not explain the points of criticism above.

AOSP is just a totally different destination, it's not a faster route to the same one.

Sailfish is spiritually MeeGo: actual Linux on the phone, not a custom skin on Google's foundations. Obviously it's faster to build a kit-car than a car factory, I don't see how that's a rebuttal, it's an entirely different conversation.

An AOSP fork on Qualcomm hardware isn't independence. Jolla are actually trying to build the factory.

The $50 fee and tablet scandal are fair hits- but fuck-ups don't make you a grifter, and we've forgiven larger players far worse.

You still haven't said what you'd actually do.

Why is AOSP a wrong path? Why would it be "tainted"? Any large enough entity can fork. Hundreds already did, successfully. Even China couldn't do otherwise - via Huawei they mutated it to HarmonyOS (becoming much different from its roots, and incompatible to it, structurally becoming superior in many ways). Why throw away 20 years of development and a sea of dev experience?

But even if you insist on a non-AOSP way: Supporting any other, more well regarded projects and initiatives? Random top of my head idea: motivate Fairphone (Denmark) to adopt some non-android OS like Ubuntu Touch?

> Why is AOSP a wrong path?

Because its existence relies on a good will of Google. See:

Google will allow only apps from verified developers to be installed on Android (9to5google.com)

https://news.ycombinator.com/item?id=45017028

and

GrapheneOS accessed Android security patches but not allowed to publish sources (grapheneos.social)

https://news.ycombinator.com/item?id=45208925

> Any large enough entity can fork.

Only megacorps will likely be able to support a hard fork for such a large codebase.

> Hundreds already did, successfully.

Which of them are hard forks? China will not be a benevolent dictator of AOSP

> Fairphone

It's Android again.

There are indeed non-Android alternatives, but not in Europe. I use Librem 5 btw.

>Because its existence relies on a good will of Google

AOSP is open source. Anyone can fork it.

>Google will allow only apps from verified developers

This is done by Play Services which is not included in AOSP even.

>Only megacorps will likely be able to support a hard fork for such a large codebase.

The same can be said about any operating system. The scope of an operating system is huge.

> The same can be said about any operating system.

GNU/Linux is already supported without a (single) megacorp. So not all OSes have this problem.

Most of Linux is written by corporations and that's just a kernel, not a full operating system like ChromeOS which took Google to be able to build.

I updated my comment to indicate GNU/Linux.

That still uses the Linux kernel which uses the work of mega corporations.

Which is not the same as one single, hostile corp.

@charcircuit

Sailfish is more like GNU/Linux, that is the OS in this context. For Jolla that is less code to maintain themselves then what Google maintains in Android/Linux. Hard forking Android/Linux looks to be quite a big bite to chew on.

Jolla doesn't need to hard fork AOSP. They can continue to benefit from the billions of dollars Google infests into AOSP and Android.

infests is a nice freudian slip.

When millions of dollars support a feature, that feature beats others- even technically superior ones, on the basis of support and polish.

We’re all playing to the tune of what Google wants because Google has the power.

Imagine a world where theres no Linux because MacOS and Windows paid lip service to people using partially functional derivatives of their OS’s, they’d still push things like liquid ass and windows recall, and those features would be spidered in.

Then people would be saying “don’t use linux, you can just use WinCore” Even though using Wincore is aiding Windows commercial interests over those of the industry as a whole.

If a product meets all of the needs of their users that's a good thing. It is a good thing to reward commercial entities for providing value to their users.

How do you know that the product meets the needs of users? From the usage? Did you hear about duopolies and walled gardens? Did you hear about banks forcing people on the duopoly?

Because if they didn't a competitor could get a foothold by solving those needs.

No, it couldn't, https://en.wikipedia.org/wiki/Vendor_lock-in

Lock in doesn't work if it doesn't solve a customers problem.

Famously, GNU/Linux solves practically all problems of normies and yet it's usage is 5%. It's not easy for them to install a new OS, and you can't buy preinstalled GNU/Linux in a store.

I don't see the issue of using AOSP. You get to skip the many years that Sailfish OS will still need in user testing. You get to skip all the possible incompatibilities with Android apps through the compatibility layer. AOSP is also Linux on the phone. I guess you mean GNU/Linux on the phone, but AOSP now also has official support for a Linux VM (you want a VM because traditional desktop Linux security is not great). They are even adding support for running Wayland apps. With the recently-added desktop support, you can plug a phone into an external screen and you'll have a desktop with Android apps and Linux desktop apps.

I think the chance of Google completely closing AOSP is pretty small, AOSP being open maintains a power equilibrium between Google and other OEMs. Closing up AOSP carries the huge risk that Samsung and some other big OEMs will fork it and Google has essentially lost the whole market overnight. I am pretty sure this is why Samsung phones also have the Galaxy Store with a bunch of apps like Netflix in it. The Galaxy Store is Samsung's subtle message to Google saying: don't try to rein us in, we can cut you out.

That said, even if Google closes AOSP, forking it and maintaining it as an open project is going to be far less work than brining Sailfish OS to the level of polish, security, etc. of AOSP.

If they can’t succeed against Google and Apple the project is useless.

Even if it’s not for lack of trying or lack of talent on their part, the fact remains their efforts could have been spent more productively elsewhere.

Reply on news.ycombinator.com